Communication networks usually comprise a plurality of transmission devices. Often, several of these devices can be used to support a service, alternatively or in combination. As the configuration of a network and the number of transmission entities within a network are subject to change during operation, fixed relations between network entities do not facilitate an optimal routing of calls.
The separation of call control and bearer control increases the number of alternatives to route a call within a network. For example in BICC (Bearer Independent Call Control) controlled networks, the call control is performed by control nodes like mobile services switching centre servers, transit switching centres, gateway mobile services switching centres, and GPRS (general packet radio system) support nodes. The payload transmission is performed by payload transmission devices like media gateways and the switching part of mobile services switching centres. Not only within but also outside of a network, nowadays a variety of alternatives exist, as several operators offer equal or compatible telecommunication services.
For the time being, calls are routed in an inflexible manner. It is state of the art that the routing of circuit switched calls is performed by using so called routing tables, comprising routing programs allocated to numbers of called subscribers. Commonly, one program is used to route all calls towards a certain subscriber number or group of subscriber numbers. Only in the case that the preferred program is not executable, e.g. due to device failure or congestion, a further program is used. To adapt the routing programs stored in the routing table, for example to a current traffic situation or changes in the network configuration, operator commands have to be performed.
This way of routing calls is quite static and forces operators either to update several routing tables at each change, or to accept a non-optimal routing of calls before an update. Non optimal routing causes costs either in network components that have to be oversized or in lack of revenues due to calls that cannot be switched.
In the U.S. Pat. No. 5,995,602, a more flexible routing scheme is described for select external networks. The patent suggests the introduction of a so-called moderator receives bids for services from external operators (carriers). The moderator sub the data, received from the carriers and often processed by said moderator, to local switches. By this, the moderator facilitates a fee minimised routing of calls from local switches through external networks.
It is a shortcoming of the solution proposed in the patent that fixed connections between carriers and the moderator are to be established. It is a further shortcoming of the invention that it only covers the connections from local switches via external networks. Another shortcoming is that the routing of calls is only optimised in respect of fees that are paid to the operators of carriers.
In the European Patent application EP 0 982 901, a bidding process is disclosed, wherein a request for bids is broadcasted for the transfer of a data packet. The received bids are analysed and a route is selected based on the analysis by using predefined metrics a cost and delay. This solution has the shortcoming that it only works in the cases that either no threshold value for a metric is defined, or that at least one bid is received that fulfils the criteria.
The British patent application GB2346507 discloses a routing method, wherein further communication elements are selected by a selection process. In the selection process, a network contacts a plurality of service providers to receive offers for a service. Based on the offers, the network selects a provider for a requested service. This solution also has the shortcoming that it only works in the cases that either no threshold value for a metric like cost or quality of service is defined, or that at least one bid is received that fulfils the criteria.